Evaluation of respondent-driven sampling in seven studies of people who use drugs from rural populations: findings from the Rural Opioid Initiative.

BACKGROUND: Accurate prevalence estimates of drug use and its harms are important to characterize burden and develop interventions to reduce negative health outcomes and disparities. Lack of a sampling frame for marginalized/stigmatized populations, including persons who use drugs (PWUD) in rural settings, makes this challenging. Respondent-driven sampling (RDS) is frequently used to recruit PWUD. However, the validity of RDS-generated population-level prevalence estimates relies on assumptions that should be evaluated. METHODS: RDS was used to recruit PWUD across seven Rural Opioid Initiative studies between 2018-2020. To evaluate RDS assumptions, we computed recruitment homophily and design effects, generated convergence and bottleneck plots, and tested for recruitment and degree differences. We compared sample proportions with three RDS-adjusted estimators (two variations of RDS-I and RDS-II) for five variables of interest (past 30-day use of heroin, fentanyl, and methamphetamine; past 6-month homelessness; and being positive for hepatitis C virus (HCV) antibody) using linear regression with robust confidence intervals. We compared regression estimates for the associations between HCV positive antibody status and (a) heroin use, (b) fentanyl use, and (c) age using RDS-1 and RDS-II probability weights and no weights using logistic and modified Poisson regression and random-effects meta-analyses. RESULTS: Among 2,842 PWUD, median age was 34 years and 43% were female. Most participants (54%) reported opioids as their drug of choice, however regional differences were present (e.g., methamphetamine range: 4-52%). Many recruitment chains were not long enough to achieve sample equilibrium. Recruitment homophily was present for some variables. Differences with respect to recruitment and degree varied across studies. Prevalence estimates varied only slightly with different RDS weighting approaches, most confidence intervals overlapped. Variations in measures of association varied little based on weighting approach. CONCLUSIONS: RDS was a useful recruitment tool for PWUD in rural settings. However, several violations of key RDS assumptions were observed which slightly impacts estimation of proportion although not associations.

Association of Alcohol Use with COVID-19 Infection and Hospitalization Among People Living with HIV in the United States, 2020.

Alcohol use was associated with elevated COVID-19 risk in the general population. People with HIV (PWH) have high prevalences of alcohol use. To evaluate the effect of alcohol use on COVID-19 risks among PWH, we estimated the risk of COVID-19 diagnosis and COVID-19-related hospitalization among PWH in routine care at 8 HIV primary care centers that contributed data to the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) cohort according to their alcohol use just prior to the COVID-19 pandemic. The CNICS data repository includes demographic characteristics, clinical diagnoses, and laboratory test results from electronic medical records and other sources. Alcohol use, substance use, and mental health symptoms were self-reported on tablet-based standardized surveys. Alcohol use was categorized according to standard, sex-specific Alcohol Use Disorder Identification Test-Consumption instrument cut-offs. We followed 5,496 PWH (79% male, 48% Black race, median age = 53 years) from March 1, 2020 to December 31, 2020. Relative to PWH with no baseline alcohol use, the adjusted hazard ratio (aHR) of COVID-19 diagnosis was 1.09 (95% confidence interval [CI]: 0.78, 1.51) for lower-risk drinking and 1.19 (95%CI: 0.81, 1.73) for unhealthy drinking. The aHR of COVID-19-related hospitalization was 0.82 (95%CI: 0.33, 1.99) for lower-risk drinking and 1.25 (95%CI: 0.50, 3.09) for unhealthy drinking. Results were not modified by recent cocaine or non-prescribed opioid use, depressive symptoms, or diagnoses of alcohol use disorder. The study suggested a slightly increased, but not statistically significant risk of COVID-19 diagnosis and hospitalization associated with unhealthy alcohol use.

Prescribing trends of proton pump inhibitors and histamine blockers among children in the United Kingdom (1998-2019): A population-based assessment.

PURPOSE: To describe the prescribing trends of proton pump inhibitors (PPIs) and H2 receptor antagonists (H2 RAs) among children with gastroesophageal reflux in the United Kingdom between 1998 and 2019. METHODS: We conducted a population-based retrospective cohort study using data from the Clinical Practice Research Datalink that included all children aged ≤18 years with a first ever diagnosis of gastroesophageal reflux between 1998 and 2019. Using negative binomial regression, we estimated crude and adjusted annual prescription rates per 1000 person-years and corresponding 95% confidence intervals (CIs) for PPIs and H2 RAs. We also assessed rate ratios of PPIs and H2 RAs prescription rates to examine changes in prescribing over time. RESULTS: Our cohort included 177 477 children with a first ever diagnosis of gastroesophageal reflux during the study period. The median age was 13 years (IQR: 1, 17) among children prescribed PPIs and 0.2 years (IQR: 0.1, 0.6) among those prescribed H2 RAs. The total prescription rate of all GERD drugs was 1468 prescriptions per 1000 person-years (PYs) (95% CI 1463-1472). Overall, PPIs had a higher prescription rate (815 per 1000 PYs, 95% CI 812-818) than H2 RAs (653 per 1000 PYs 95% CI 650-655). Sex- and age-adjusted rate ratios of 2019 versus 1998 demonstrated a 10% increase and a 76% decrease in the prescription rates of PPIs and H2 RAs, respectively. CONCLUSIONS: Prescription rates for PPIs increased, especially during the first half of the study period, while prescription rates for H2 RA decreased over time.

Retrospective Cohort Studies in Craniofacial Outcomes Research: An Epidemiologist's Approach to Mitigating Bias.

Retrospective cohort studies, defined as a follow-up study in which outcome occurred prior to study onset, are common in craniofacial outcomes research and will continue to be prevalent given the increasing availability of secondary datasets and inherent prospective study limitations. However, if available data are not adequately measured, or necessary variables are absent, retrospective cohort studies can be particularly prone to bias. This brief communication aims to highlight the primary sources of bias, including measurement error, selection bias, and confounding. Each source is clearly defined, examples pertinent to craniofacial outcomes are provided, and mitigation strategies are discussed.

Associations between alcohol and cigarette use and type 1 and 2 myocardial infarction among people with HIV.

OBJECTIVES: People with HIV have a higher risk of myocardial infarction (MI) than the general population, with a greater proportion of type 2 MI (T2MI) due to oxygen demand-supply mismatch compared with type 1 (T1MI) resulting from atherothrombotic plaque disruption. People living with HIV report a greater prevalence of cigarette and alcohol use than do the general population. Alcohol use and smoking as risk factors for MI by type are not well studied among people living with HIV. We examined longitudinal associations between smoking and alcohol use patterns and MI by type among people living with HIV. DESIGN AND METHODS: Using longitudinal data from the Centers for AIDS Research Network of Integrated Clinical Systems cohort, we conducted time-updated Cox proportional hazards models to determine the impact of smoking and alcohol consumption on adjudicated T1MI and T2MI. RESULTS: Among 13 506 people living with HIV, with a median 4 years of follow-up, we observed 177 T1MI and 141 T2MI. Current smoking was associated with a 60% increase in risk of both T1MI and T2MI. In addition, every cigarette smoked per day was associated with a 4% increase in risk of T1MI, with a suggestive, but not significant, 2% increase for T2MI. Cigarette use had a greater impact on T1MI for men than for women and on T2MI for women than for men. Increasing alcohol use was associated with a lower risk of T1MI but not T2MI. Frequency of heavy episodic alcohol use was not associated with MI. CONCLUSIONS: Our findings reinforce the prioritization of smoking reduction, even without cessation, and cessation among people living with HIV for MI prevention and highlight the different impacts on MI type by gender.

Association of Immune Cell Subsets with Incident Hip Fracture: The Cardiovascular Health Study.

In this study, we aimed to evaluate the association of innate and adaptive immune cell subsets in peripheral blood mononuclear cells (PBMCs) with hip fracture. To conduct this study, we used data from the Cardiovascular Health Study (CHS), a U.S. multicenter observational cohort of community-dwelling men and women aged ≥ 65 years. Twenty-five immune cell phenotypes were measured by flow cytometry from cryopreserved PBMCs of CHS participants collected in 1998-1999. The natural killer (NK), γδ T, T helper 17 (Th17), and differentiated/senescent CD4+CD28- T cell subsets were pre-specified as primary subsets of interest. Hip fracture incidence was assessed prospectively by review of hospitalization records. Multivariable Cox hazard models evaluated associations of immune cell phenotypes with incident hip fracture in sex-stratified and combined analyses. Among 1928 persons, 259 hip fractures occurred over a median 9.7 years of follow-up. In women, NK cells were inversely associated with hip fracture [hazard ratio (HR) 0.77, 95% confidence interval (CI) 0.60-0.99 per one standard deviation higher value] and Th17 cells were positively associated with hip fracture [HR 1.18, 95% CI 1.01-1.39]. In men, γδ T cells were inversely associated with hip fracture [HR 0.60, 95% CI 0.37-0.98]. None of the measured immune cell phenotypes were significantly associated with hip fracture incidence in combined analyses. In this large prospective cohort of older adults, potentially important sex differences in the associations of immune cell phenotypes and hip fracture were identified. However, immune cell phenotypes had no association with hip fracture in analyses combining men and women.

Developing a prediction model of children asthma risk using population-based family history health records.

BACKGROUND: Identifying children at high risk of developing asthma can facilitate prevention and early management strategies. We developed a prediction model of children's asthma risk using objectively collected population-based children and parental histories of comorbidities. METHODS: We conducted a retrospective population-based cohort study using administrative data from Manitoba, Canada, and included children born from 1974 to 2000 with linkages to ≥1 parent. We identified asthma and prior comorbid condition diagnoses from hospital and outpatient records. We used two machine-learning models: least absolute shrinkage and selection operator (LASSO) logistic regression (LR) and random forest (RF) to identify important predictors. The predictors in the base model included children's demographics, allergic conditions, respiratory infections, and parental asthma. Subsequent models included additional multiple comorbidities for children and parents. RESULTS: The cohort included 195,666 children: 51.3% were males and 17.7% had asthma diagnosis. The base LR model achieved a low predictive performance with sensitivity of 0.47, 95% confidence interval (0.45-0.48), and specificity of 0.67 (0.66-0.67) using a predicted probability threshold of 0.20. Sensitivity significantly improved when children's comorbidities were included using LASSO LR: 0.71 (0.69-0.72). Predictive performance further improved by including parental comorbidities (sensitivity = 0.72 [0.70-0.73], specificity = 0.69 [0.69-0.70]). We observed similar results for the RF models. Children's menstrual disorders and mood and anxiety disorders, parental lipid metabolism disorders and asthma were among the most important variables that predicted asthma risk. CONCLUSION: Including children and parental comorbidities to children's asthma prediction models improves their accuracy.

Clinical and Behavioral Outcomes for Transgender Women Engaged in HIV Care: Comparisons to Cisgender Men and Women in the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) Cohort.

Describe health of transgender women (TW) with HIV vs. cisgender men and women (CM, CW) in a U.S. HIV care cohort. Data were from Centers for AIDS Research Network of Integrated Clinical Systems (CNICS), 2005-2022. TW were identified using clinical data/identity measures. PWH (n = 1285) were included in analyses (275 TW, 547 CM, 463 CW). Cross-sectional multivariable analyses compared HIV outcomes/co-morbidities between TW/CM and TW/CW, and adjusted odds ratios (aOR) and 95% confidence intervals (95% CI) were estimated. TW had poorer adherence (> 90% adherent; aOR 0.57; 95%CI 0.38, 0.87) and were more likely to miss ≥ 3 visits in the past year than CM (aOR 1.50, 95%CI 1.06, 2.10); indicated more anxiety compared to both CM and CW (p ≤ 0.001, p = 0.02); hepatitis C infection (p = 0.03) and past-year/lifetime substance treatment (p = 0.004/p = 0.001) compared to CM; and substance use relative to CW. TW with HIV differed in HIV clinical outcomes and co-morbidities from CM and CW.

Current Antiretroviral Treatment Among People With Human Immunodeficiency Virus in the United States: Findings from the Centers for AIDS Research Network of Integrated Clinic Systems Cohort.

Among 14 049 people with human immunodeficiency virus in care in 2019-2020, 96% were treated with antiretroviral therapy (ART). Current antiretroviral treatment patterns highlight high uptake of guideline-recommended ART regimens including second-generation integrase strand transfer inhibitors (dolutegravir and bictegravir) and tenofovir alafenamide, especially in antiretroviral-naive individuals initiating ART.

Tobacco smoking and binge alcohol use are associated with incident venous thromboembolism in an HIV cohort.

BACKGROUND: People with HIV (PWH) are at increased risk of cardiovascular comorbidities and substance use is a potential predisposing factor. We evaluated associations of tobacco smoking and alcohol use with venous thromboembolism (VTE) in PWH. METHODS: We assessed incident, centrally adjudicated VTE among 12 957 PWH within the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) cohort between January 2009 and December 2018. Using separate Cox proportional hazards models, we evaluated associations of time-updated alcohol and cigarette use with VTE, adjusting for demographic and clinical characteristics. Smoking was evaluated as pack-years and never, former, or current use with current cigarettes per day. Alcohol use was parameterized using categorical and continuous alcohol use score, frequency of use, and binge frequency. RESULTS: During a median of 3.6 years of follow-up, 213 PWH developed a VTE. One-third of PWH reported binge drinking and 40% reported currently smoking. In adjusted analyses, risk of VTE was increased among both current (HR: 1.44, 95% CI: 1.02-2.03) and former (HR: 1.44, 95% CI: 0.99-2.07) smokers compared to PWH who never smoked. Additionally, total pack-years among ever-smokers (HR: 1.10 per 5 pack-years; 95% CI: 1.03-1.18) was associated with incident VTE in a dose-dependent manner. Frequency of binge drinking was associated with incident VTE (HR: 1.30 per 7 days/month, 95% CI: 1.11-1.52); however, alcohol use frequency was not. Severity of alcohol use was not significantly associated with VTE. CONCLUSIONS: Current smoking and pack-year smoking history were dose-dependently associated with incident VTE among PWH in CNICS. Binge drinking was also associated with VTE. Interventions for smoking and binge drinking may decrease VTE risk among PWH.

Nonclassical Monocytes (CD14dimCD16+) Are Associated With Carotid Intima-Media Thickness Progression for Men but Not Women: The Multi-Ethnic Study of Atherosclerosis-Brief Report.

[Figure: see text].

The Use of Active Comparators in Self-Controlled Designs.

For self-controlled studies of medication-related effects, time-varying confounding by indication can occur if the indication varies over time. We describe how active comparators might mitigate such bias, using an empirical example. Approaches to using active comparators are described for case-crossover design, case-time-control design, self-controlled case-series, and sequence symmetry analyses. In the empirical example, we used Danish data from 1996-2018 to study the association between penicillin and venous thromboembolism (VTE), using roxithromycin, a macrolide antibiotic, as comparator. Upper respiratory infection is a transient risk factor for VTE, thus representing time-dependent confounding by indication. Odds ratios for case-crossover analysis were 3.35 (95% confidence interval: 3.23, 3.49) for penicillin and 3.56 (95% confidence interval: 3.30, 3.83) for roxithromycin. We used a Wald-based method or an interaction term to estimate the odds ratio for penicillin with roxithromycin as comparator. These 2 estimates were 0.94 (95% confidence interval: 0.87, 1.03) and 1.03 (95% confidence interval: 0.95, 1.13). Results were similar for the case-time-control analysis, but both the self-controlled case-series and sequence symmetry analysis suggested a weak protective effect of penicillin, seemingly explained by VTE affecting future exposure exclusively for penicillin. The strong association of antibiotics with VTE suggests presence of confounding by indication. Such confounding can be mitigated by using an active comparator.

Changing Patterns of Alcohol Use and Probability of Unsuppressed Viral Load Among Treated Patients with HIV Engaged in Routine Care in the United States.

We examined HIV viral load non-suppression ([Formula: see text] 200 copies/mL) subsequent to person-periods (3-18 months) bookended by two self-reports of alcohol use on a standardized patient reported outcome assessment among adults in routine HIV care. We examined the relative risk (RR) of non-suppression associated with increases and decreases in alcohol use (relative to stable use), stratified by use at the start of the person-period. Increases in drinking from abstinence were associated with higher risk of viral non-suppression (low-risk without binge: RR 1.16, 95% CI 1.03, 1.32; low-risk with binge: RR 1.35, 95% CI 1.11, 1.63; high-risk: RR 1.89, 95% CI 1.16, 3.08). Decreases in drinking from high-risk drinking were weakly, and not statistically significantly associated with lower risk of viral non-suppression. Other changes in alcohol use were not associated with viral load non-suppression. Most changes in alcohol consumption among people using alcohol at baseline were not strongly associated with viral non-suppression.

Drug and alcohol use among people living with HIV in care in the United States by geographic region.

Substance use in the U.S. varies by geographic region. Opioid prescribing practices and marijuana, heroin, and methamphetamine availability are evolving differently across regions. We examined self-reported substance use among people living with HIV (PLWH) in care at seven sites from 2017-2019 to understand current regional substance use patterns. We calculated the percentage and standardized percentage of PLWH reporting current drug use and at-risk and binge alcohol use by U.S. Census Bureau geographic region and examined associations in adjusted logistic regression analyses. Among 7,686 PLWH, marijuana use was the most prevalent drug (30%), followed by methamphetamine/crystal (8%), cocaine/crack (7%), and illicit opioids (3%). One-third reported binge alcohol use (32%). Differences in percent of current use by region were seen for marijuana (24-41%) and methamphetamine/crystal (2-15%), with more use in the West and Northeast, and binge alcohol use (26-40%). In adjusted analyses, PLWH in the Midwest were significantly less likely to use methamphetamine/crystal (aOR: 0.13;0.06-0.25) or illicit opioids (aOR:0.16;0.05-0.53), and PLWH in the Northeast were more likely to use cocaine/crack (aOR:1.59;1.16-2.17), compared to PLWH in the West. Understanding differences in substance use patterns in the current era, as policies continue to evolve, will enable more targeted interventions in clinical settings among PLWH.

Natural killer cells, gamma delta T cells and classical monocytes are associated with systolic blood pressure in the multi-ethnic study of atherosclerosis (MESA).

BACKGROUND: Hypertension is a major source of cardiovascular morbidity and mortality. Recent evidence from mouse models, genetic, and cross-sectional human studies suggest increased proportions of selected immune cell subsets may be associated with levels of systolic blood pressure (SBP). METHODS: We assayed immune cells from cryopreserved samples collected at the baseline examination (2000-2002) from 1195 participants from the multi-ethnic study of atherosclerosis (MESA). We used linear mixed models, with adjustment for age, sex, race/ethnicity, smoking, exercise, body mass index, education, diabetes, and cytomegalovirus titers, to estimate the associations between 30 immune cell subsets (4 of which were a priori hypotheses) and repeated measures of SBP (baseline and up to four follow-up measures) over 10 years. The analysis provides estimates of the association with blood pressure level. RESULTS: The mean age of the MESA participants at baseline was 64 ± 10 years and 53% were male. A one standard deviation (1-SD) increment in the proportion of γδ T cells was associated with 2.40 mmHg [95% confidence interval (CI) 1.34-3.42] higher average systolic blood pressure; and for natural killer cells, a 1-SD increment was associated with 1.88 mmHg (95% CI 0.82-2.94) higher average level of systolic blood pressure. A 1-SD increment in classical monocytes (CD14++CD16-) was associated with 2.01 mmHG (95% CI 0.79-3.24) lower average systolic blood pressure. There were no associations of CD4+ T helper cell subsets with average systolic blood pressure. CONCLUSION: These findings suggest that the innate immune system plays a role in levels of SBP whereas there were no associations with adaptive immune cells.

Impact of Abstinence and of Reducing Illicit Drug Use Without Abstinence on Human Immunodeficiency Virus Viral Load.

BACKGROUND: Substance use is common among people living with human immunodeficiency virus (PLWH) and a barrier to achieving viral suppression. Among PLWH who report illicit drug use, we evaluated associations between HIV viral load (VL) and reduced use of illicit opioids, methamphetamine/crystal, cocaine/crack, and marijuana, regardless of whether or not abstinence was achieved. METHODS: This was a longitudinal cohort study of PLWH from 7 HIV clinics or 4 clinical studies. We used joint longitudinal and survival models to examine the impact of decreasing drug use and of abstinence for each drug on viral suppression. We repeated analyses using linear mixed models to examine associations between change in frequency of drug use and VL. RESULTS: The number of PLWH who were using each drug at baseline ranged from n = 568 (illicit opioids) to n = 4272 (marijuana). Abstinence was associated with higher odds of viral suppression (odds ratio [OR], 1.4-2.2) and lower relative VL (ranging from 21% to 42% by drug) for all 4 drug categories. Reducing frequency of illicit opioid or methamphetamine/crystal use without abstinence was associated with VL suppression (OR, 2.2, 1.6, respectively). Reducing frequency of illicit opioid or methamphetamine/crystal use without abstinence was associated with lower relative VL (47%, 38%, respectively). CONCLUSIONS: Abstinence was associated with viral suppression. In addition, reducing use of illicit opioids or methamphetamine/crystal, even without abstinence, was also associated with viral suppression. Our findings highlight the impact of reducing substance use, even when abstinence is not achieved, and the potential benefits of medications, behavioral interventions, and harm-reduction interventions.

Association Between Chronic Hepatitis C Virus Infection and Myocardial Infarction Among People Living With HIV in the United States.

Hepatitis C virus (HCV) infection is common among people living with human immunodeficiency virus (PLWH). Extrahepatic manifestations of HCV, including myocardial infarction (MI), are a topic of active research. MI is classified into types, predominantly atheroembolic type 1 MI (T1MI) and supply-demand mismatch type 2 MI (T2MI). We examined the association between HCV and MI among patients in the Centers for AIDS Research (CFAR) Network of Integrated Clinical Systems, a US multicenter clinical cohort of PLWH. MIs were centrally adjudicated and categorized by type using the Third Universal Definition of Myocardial Infarction. We estimated the association between chronic HCV (RNA+) and time to MI while adjusting for demographic characteristics, cardiovascular risk factors, clinical characteristics, and history of injecting drug use. Among 23,407 PLWH aged ≥18 years, there were 336 T1MIs and 330 T2MIs during a median of 4.7 years of follow-up between 1998 and 2016. HCV was associated with a 46% greater risk of T2MI (adjusted hazard ratio (aHR) = 1.46, 95% confidence interval (CI): 1.09, 1.97) but not T1MI (aHR = 0.87, 95% CI: 0.58, 1.29). In an exploratory cause-specific analysis of T2MI, HCV was associated with a 2-fold greater risk of T2MI attributed to sepsis (aHR = 2.01, 95% CI: 1.25, 3.24). Extrahepatic manifestations of HCV in this high-risk population are an important area for continued research.

Genetic architecture of cardiometabolic risks in people living with HIV.

BACKGROUND: Advances in antiretroviral therapies have greatly improved the survival of people living with human immunodeficiency virus (HIV) infection (PLWH); yet, PLWH have a higher risk of cardiovascular disease than those without HIV. While numerous genetic loci have been linked to cardiometabolic risk in the general population, genetic predictors of the excessive risk in PLWH are largely unknown. METHODS: We screened for common and HIV-specific genetic variants associated with variation in lipid levels in 6284 PLWH (3095 European Americans [EA] and 3189 African Americans [AA]), from the Centers for AIDS Research Network of Integrated Clinical Systems cohort. Genetic hits found exclusively in the PLWH cohort were tested for association with other traits. We then assessed the predictive value of a series of polygenic risk scores (PRS) recapitulating the genetic burden for lipid levels, type 2 diabetes (T2D), and myocardial infarction (MI) in EA and AA PLWH. RESULTS: We confirmed the impact of previously reported lipid-related susceptibility loci in PLWH. Furthermore, we identified PLWH-specific variants in genes involved in immune cell regulation and previously linked to HIV control, body composition, smoking, and alcohol consumption. Moreover, PLWH at the top of European-based PRS for T2D distribution demonstrated a > 2-fold increased risk of T2D compared to the remaining 95% in EA PLWH but to a much lesser degree in AA. Importantly, while PRS for MI was not predictive of MI risk in AA PLWH, multiethnic PRS significantly improved risk stratification for T2D and MI. CONCLUSIONS: Our findings suggest that genetic loci involved in the regulation of the immune system and predisposition to risky behaviors contribute to dyslipidemia in the presence of HIV infection. Moreover, we demonstrate the utility of the European-based and multiethnic PRS for stratification of PLWH at a high risk of cardiometabolic diseases who may benefit from preventive therapies.

Incorporating sampling weights into robust estimation of Cox proportional hazards regression model, with illustration in the Multi-Ethnic Study of Atherosclerosis.

BACKGROUND: Cox proportional hazards regression models are used to evaluate associations between exposures of interest and time-to-event outcomes in observational data. When exposures are measured on only a sample of participants, as they are in a case-cohort design, the sampling weights must be incorporated into the regression model to obtain unbiased estimating equations. METHODS: Robust Cox methods have been developed to better estimate associations when there are influential outliers in the exposure of interest, but these robust methods do not incorporate sampling weights. In this paper, we extend these robust methods, which already incorporate influence weights, so that they also accommodate sampling weights. RESULTS: Simulations illustrate that in the presence of influential outliers, the association estimate from the weighted robust method is closer to the true value than the estimate from traditional weighted Cox regression. As expected, in the absence of outliers, the use of robust methods yields a small loss of efficiency. Using data from a case-cohort study that is nested within the Multi-Ethnic Study of Atherosclerosis (MESA) longitudinal cohort study, we illustrate differences between traditional and robust weighted Cox association estimates for the relationships between immune cell traits and risk of stroke. CONCLUSIONS: Robust weighted Cox regression methods are a new tool to analyze time-to-event data with sampling, e.g. case-cohort data, when exposures of interest contain outliers.

Association of Aspirin and Other Nonsteroidal Anti-inflammatory Drugs With Vertebral Trabecular Bone: Data From Multiethnic Study of Atherosclerosis, a Population-Based Multicenter Cohort Study.

OBJECTIVE: The objective of this article was to study the association of aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) with bone mineral density (BMD). METHODS: Spine BMD was evaluated in a subset of 2028 participants from the Multiethnic Study of Atherosclerosis cohort who were NSAID users (including aspirin) and underwent both lumbar and thoracic imaging. Multiethnic Study of Atherosclerosis is a prospective cohort study that includes 4 ethnic groups (white, Asian, African American, and Hispanic). Trabecular BMD was evaluated by quantitative computed tomography based on cardiac computed tomography images, which were obtained during coronary calcium scans. The analyses were cross sectional using baseline examination data for exposure and outcomes. RESULTS: After adjustment for potential confounders including age, sex, race, and traditional cardiovascular risk factors, a small association between trabecular BMD and baseline use of COX-2-selective NSAID was observed. COX-2-selective NSAID use was associated with 7.4 mg/cm (95% confidence interval [CI], 1.6-13.3; P = 0. 013) higher trabecular BMD in thoracic spine and 10.6 mg/cm higher at lumbar spine (95% CI, 5.1-16.1; P < 0.001). Among regular aspirin users, there was no association between drug use and trabecular BMD. Considering all spine fractures together, the prevalence ratio of fractures among aspirin users was 1.0 (95% CI, 0.6-1.6) and 1.1 (95% CI, 0.5-2.3) among COX-2-selective NSAID users. CONCLUSIONS: Regular use of aspirin has no significant association with trabecular BMD in either the thoracic or lumbar spine and no association with fracture prevalence. COX-2-selective NSAIDs may have modest positive association with BMD, but the mechanisms were not assessed and the observational study design makes residual confounding a possible alternate explanation. Potential pathological mechanisms warrant further longitudinal exploration.